| [10:11:50] | * laprice has joined #aegir |
| [10:29:44] | * gandhiano_ has quit (Remote host closed the connection) |
| [10:39:49] | * omega8cc_ has joined #aegir |
| [10:41:40] | * omega8cc has quit (Ping timeout: 260 seconds) |
| [10:41:40] | * omega8cc_ is now known as omega8cc |
| [11:09:55] | * realityloop has joined #aegir |
| [11:11:29] | <realityloop> | omega8cc: is it possible to reduce the frequency of which reports are sent for OS updates on a BOA,Barracuda setup? |
| [11:11:29] | <hefring> | realityloop: 12 hours 27 min ago <omega8cc> tell realityloop: no, unless you are on some crappy system like virtuozzo family. On my test linode, with 1GB ram it is using stable 400MB plus around 100MB swap after one week of testing all available Octopus platforms, so there are many sites created etc. |
| [11:16:29] | <omega8cc> | realityloop: no, but feel free to open a feature request |
| [11:17:30] | <realityloop> | which queue is best to open it in? |
| [11:18:13] | * kylemathews has quit (Read error: Operation timed out) |
| [11:18:28] | <omega8cc> | realityloop: barracuda |
| [11:20:54] | <realityloop> | omega8cc: done, http://drupal.org/node/1653410 |
| [11:20:55] | <hefring> | http://drupal.org/node/1653410 => Is it possible to set how often the reports about OS updates are emailled out? => Barracuda, Miscellaneous, normal, active, 0 comments, 1 IRC mention |
| [11:22:48] | <realityloop> | omega8cc: also thanks for the reply on memory usage.. it appears that urpad use crappy system as box is using most of 1.5GB with no sites running.. |
| [11:22:59] | <realityloop> | gotta go, later |
| [11:24:15] | * realityloop has quit (Remote host closed the connection) |
| [11:45:30] | * Pomliane has quit (Quit: Pomliane) |
| [12:04:20] | * grobot has quit (Quit: ^d) |
| [12:10:19] | * patcon has joined #aegir |
| [12:20:14] | * gusaus has quit (Quit: gusaus) |
| [12:29:01] | * Pomliane has joined #aegir |
| [12:29:28] | * Pomliane has quit (Client Quit) |
| [13:07:18] | * gboudrias has quit (Quit: Leaving.) |
| [13:54:33] | * AquaticDisorder has joined #aegir |
| [14:06:20] | <jasonmce> | just want to confirm that I shouldn't "aptitude full-upgrade" my BOA box, but instead should wget the latest BOA.sh.txt and run it in a bash shell, then "barracuda up-stable" |
| [14:06:47] | <jasonmce> | aegir-jenkins ping |
| [14:08:21] | * patcon has quit (Ping timeout: 248 seconds) |
| [14:27:58] | * kylemathews has joined #aegir |
| [15:31:35] | * banghouse has quit (Remote host closed the connection) |
| [15:33:55] | * EclipseGc has quit (Ping timeout: 260 seconds) |
| [16:22:54] | * Egyptian[Laptop] has quit (Ping timeout: 272 seconds) |
| [16:28:27] | * Blegh has quit (Ping timeout: 246 seconds) |
| [16:30:24] | * Blegh has joined #aegir |
| [17:20:33] | * kylemathews has quit (Ping timeout: 255 seconds) |
| [17:58:39] | * gusaus has joined #aegir |
| [18:24:50] | * gusaus has quit (Quit: gusaus) |
| [18:48:10] | * bgm has quit (K-Lined) |
| [18:54:32] | * ivanjaros has joined #aegir |
| [19:13:03] | * svendecabooter has joined #aegir |
| [20:41:09] | * eft has quit (Ping timeout: 244 seconds) |
| [22:54:06] | * EclipseGc has joined #aegir |
| [23:40:40] | * svendecabooter has quit (Remote host closed the connection) |
| [00:05:39] | * EclipseGc has quit (Quit: EclipseGc) |
| [00:13:56] | * Erobaka has quit (Quit: Client Aion _[RSW]_ A bientot :) |
| [00:31:43] | * patcon has joined #aegir |
| [00:31:53] | * EclipseGc has joined #aegir |
| [00:38:10] | * patcon has quit (Ping timeout: 260 seconds) |
| [00:41:18] | * fatguylaughing has joined #aegir |
| [00:43:33] | * rv0 has joined #aegir |
| [00:51:39] | * zombiebeard has joined #aegir |
| [01:22:56] | * banghouse has joined #aegir |
| [01:26:45] | * mrfelton has joined #aegir |
| [01:37:15] | * gandhiano has joined #aegir |
| [02:12:00] | * kylemathews has joined #aegir |
| [02:25:00] | * Pomliane has joined #aegir |
| [02:45:38] | * EclipseGc has quit (Quit: EclipseGc) |
| [03:36:55] | * gandhiano has quit (Remote host closed the connection) |
| [03:37:17] | * gusaus has joined #aegir |
| [03:39:09] | * gusaus has quit (Client Quit) |
| [03:40:12] | * gusaus has joined #aegir |
| [03:43:22] | * rv0 has quit (Quit: rv0) |
| [04:10:25] | * eft has joined #aegir |
| [04:12:10] | * gusaus has quit (Quit: gusaus) |
| [04:16:30] | * ivanjaros has quit (Quit: Miranda IM! Smaller, Faster, Easier. http://miranda-im.org) |
| [04:21:12] | * zombiebeard has quit (Quit: zombiebeard) |
| [04:23:04] | * gusaus has joined #aegir |
| [04:23:26] | * gusaus has quit (Read error: Connection reset by peer) |
| [04:23:45] | * gusaus has joined #aegir |
| [04:30:24] | * fatguylaughing has quit (Quit: fatguylaughing) |
| [04:49:44] | * wizonesolutions has quit (Ping timeout: 245 seconds) |
| [05:14:32] | * rv0 has joined #aegir |
| [05:15:46] | * rv0 has quit (Client Quit) |
| [05:16:13] | * fatguylaughing has joined #aegir |
| [05:28:23] | * fatguylaughing has quit (Quit: fatguylaughing) |
| [05:30:05] | * laprice has left #aegir () |
| [05:31:50] | * gusaus has quit (Quit: gusaus) |
| [05:43:09] | * wizonesolutions has joined #aegir |
| [06:33:49] | * lsolesen has joined #aegir |
| [06:36:08] | <lsolesen> | I have been running aegir on a linode server for some time now, and it works great on just a regular LAMP stack with apache and MySQL. Reading a little more about it I stumbled onto barracuda and octopus. As I understand it barracuda is just an install script, while octopus makes it possible to create mini-aegir instances so it is easier to give a developer access to one site without giving him access to other sites. |
| [06:37:10] | <lsolesen> | Should I upgrade my server to use barracuda and octopus. Should I just use octopus, and is it easy to transfer sites to an octopus instance, if I choose so? |
| [07:29:59] | <mrfelton> | hmm.. Barracuda upgrade has been running for 2.5 hours now. That's not right |
| [07:30:02] | <mrfelton> | what to do?! |
| [07:31:17] | <mrfelton> | lsolesen: you can't really upgrade your existing server to Barracuda - You have to start with a a completely fresh server to use Barracuda. |
| [07:31:36] | <lsolesen> | mrfelton: Sorry, I do not know the answer to your question. Why are you using Barracuda instead of just a regular LAMP-stack? |
| [07:33:10] | <lsolesen> | mrfelton: do you need barracuda to do octopus? |
| [07:33:40] | <lsolesen> | mrfelton: you cannot it seems. |
| [07:33:46] | <mrfelton> | lsolesen: We don't always use Barracuda. Depends on the needs of the client etc. But generally, Barracuda provides a very robust and well thought out hosting envirionment, specifically designed to serve Drupal sites in conjunction with Aegir, and lots of great stuff for performance configured nicer. |
| [07:34:12] | <mrfelton> | lsolesen: you do need barracuda to use octopus. But you don't need octopus to use barracuda |
| [07:34:29] | <lsolesen> | Hm. Ok. Should have read more about it before I setup my servers ... |
| [07:34:50] | <lsolesen> | have two different aegir servers for different clients. |
| [07:35:03] | <lsolesen> | Should have just put barracuda on both of them. |
| [07:35:06] | <lsolesen> | :/ |
| [07:36:01] | <lsolesen> | maybe I could easily transfer all sites from one aegir instance to the other aegir instance and reinstall the server... will investigate that. |
| [07:36:16] | <lsolesen> | reinstall one server at the time. |
| [07:36:50] | <mrfelton> | lsolesen: It's doable. Check omega8cc's website. There is some good info there about importing sites in to Aegir |
| [07:37:46] | * patcon has joined #aegir |
| [07:40:18] | * patcon has quit (Read error: Connection reset by peer) |
| [07:40:22] | <lsolesen> | Cool: Found this also: http://netengine.com.au/blog/moving-drupal-sites-between-aegir-instances |
| [07:41:54] | * patcon has joined #aegir |
| [07:50:52] | <mrfelton> | lsolesen: that is a little bit out of date actually. http://community.aegirproject.org/content/content/administrator/post-ins... is a better one |
| [07:51:33] | <lsolesen> | mrfelton: ok. thanks. |
| [07:51:38] | <lsolesen> | See when I have time. |
| [07:51:50] | <lsolesen> | Do not break a working environment unless you have to :) |
| [07:52:31] | <lsolesen> | mrfelton: does aegir need to be able to sudo -- it seems it does to use apt-get to upgrade the debian package to the newest aegir. |
| [07:53:18] | <mrfelton> | lsolesen: the aegir user doesn't. Aegir doesn't update anything for you. You have to log in as root to update Barracuda |
| [07:53:41] | <lsolesen> | mrfelton: I do not have barracuda. just aegir. |
| [07:53:58] | <lsolesen> | mrfelton: it just tell me that it expects me to run the script as aegir. |
| [07:54:13] | <mrfelton> | lsolesen: ok, we still aegir will not update anything for you. What script are you talking about? |
| [07:54:14] | <lsolesen> | when i do apt-get install aegir to upgrade the aegir instance on my box. |
| [07:54:46] | <lsolesen> | mrfelton: http://community.aegirproject.org/upgrading/debian |
| [07:54:51] | <mrfelton> | lsolesen: I have never actually installed aegir from apt. But, you shouldn't give aegir sudo. You should log in as another user to perform that task. |
| [07:55:07] | <mrfelton> | a user that has sudo. O, root |
| [07:55:11] | <mrfelton> | or |
| [07:55:47] | <lsolesen> | mrfelton: strange. Am a little reluctant when the apt-get actually tells me about the assumptions for the package install.. |
| [07:56:20] | <mrfelton> | lsolesen: yeah, you should log in under your own account on the server, which should probably have sudo access. And, you would need to run those commands with sudo. Or log in as root (not recommended!) |
| [07:56:58] | <mrfelton> | apt-get is a pretty solid packaging system. I wouldn't be too concerned about dependencies that it lists or anything like that |
| [07:58:12] | <mrfelton> | ok, so I fixed my update problem. Trick is to tail the update log in another terminal window whist it runs. |
| [07:59:27] | <mrfelton> | omega8cc: problem was that the new-relic key wasn't in place, so it was prompting me to accept the untrusted source. But the Barracuda installer hid this from me. Shouldn't it install the key for me? I stopped the upgrade, installed the key manually, and ran it again and now its past that |
| [08:02:39] | <omega8cc> | mrfelton: it was upgrade on server with new relic installed already? |
| [08:02:51] | <mrfelton> | omega8cc: it was, yes |
| [08:03:03] | <omega8cc> | but with boa or manually? |
| [08:03:13] | <mrfelton> | omega8cc: good point. It was installed manually |
| [08:04:25] | <mrfelton> | Still, I know the rule is that you shouldn't install stuff manually on barracuda servers. But I would suspect that quite a few people have installed new relic on their barracuda servers, since it wasn't there previously |
| [08:04:30] | <omega8cc> | yeah, boa is not smart enough to detect stuff installed under or above its radar |
| [08:06:18] | <mrfelton> | Mind you, I probably should have installed the key when I did that anyway :p |
| [08:06:38] | <omega8cc> | it is rather hard to support non-standard stuff already installed, as boa is not aware of the way it was done etc, so it would need to grep /etc/apt/sources.list.d/* etc |
| [08:07:28] | <omega8cc> | yeah, you should ;) |
| [08:08:08] | <omega8cc> | boa adds the key only if /etc/newrelic/newrelic.cfg doesn't exist |
| [08:08:28] | <mrfelton> | right |
| [08:10:59] | <mrfelton> | Hey, wonder if you can help with something actually. I'm having some real trouble with ssl<->non-ssl redirects on several of our aegir/barracuda servers. Tried with secure pages, as well as the method in the provided sample override.global.inc file, and in all cases we get problems with cookies not being shared properly across http/https. Ever experienced this? I believe it might be related to a core issue or limitation where i |
| [08:10:59] | <mrfelton> | not possible to upgrade sessions from http to https. So if your session starts in http, and then you switch to https, thing get lost.. any ideas?! |
| [08:11:45] | <mrfelton> | One one site, I've had to resort to forcing the entire site into https, which gets around the problem. |
| [08:12:24] | <mrfelton> | But on other sites, we can't do that because some of the external resources we are using are only available over http, so those parts of the site need to be http in order to avoid the insecure content warnings from browsers |
| [08:15:59] | <mrfelton> | omega8cc: I'm pretty sure that the problem is related to the micro caching feature in some way |
| [08:16:01] | * gusaus has joined #aegir |
| [08:16:48] | <omega8cc> | not |
| [08:17:24] | <omega8cc> | it is drupal |
| [08:17:33] | <omega8cc> | it is well known issue |
| [08:17:59] | <mrfelton> | omega8cc: it is drupal, ok. Its the cookie upgrading issue right? |
| [08:18:07] | <omega8cc> | you need to use workarounds as it is pretty much broken in the drupal core |
| [08:18:09] | <mrfelton> | do you know of a workaround? |
| [08:18:14] | <mrfelton> | It's just killing us at the moment |
| [08:18:15] | <mrfelton> | ] |
| [08:19:14] | <omega8cc> | maybe check drupal core patches listed/discussed here: http://drupal.org/project/securepages and start from there |
| [08:19:31] | <mrfelton> | I found one workaround for secure pages, which was supposed to forcefully upgrade the session. But it doesn't really seem to work. |
| [08:19:32] | <mrfelton> | http://drupal.org/node/1545970 |
| [08:19:33] | <hefring> | http://drupal.org/node/1545970 => if ssid is empty on a session, a new session is generated. => Secure Pages, Code, major, needs review, 10 comments, 1 IRC mention |
| [08:20:00] | <mrfelton> | omega8cc: I've been through all these patches. Nothing seems to do it |
| [08:22:45] | <omega8cc> | usually we are using our own logic, adjusted per site, in the global.inc to handle this, so user keeps the session, but I didn't work on this personally for a long time already, so feel free to open a support request in the octopus or barracuda queue and maybe Robert or Greg will be able to share some how-to |
| [08:23:33] | <omega8cc> | mrfelton: ^^ |
| [08:24:57] | <omega8cc> | mrfelton: also, Secure Pages module is just another layer of the problem, not a solution |
| [08:25:16] | <mrfelton> | omega8cc: fair enough. We too have a lot of custom handling in global.inc, but nothing that has managed to ensure that cookies can consistently get upgraded |
| [08:25:57] | <mrfelton> | omega8cc: yeah, we have come to the same conclusion. Securepages is not needed, and we are not using it. We can do it in global.inc. Mostly. |
| [08:26:19] | <mrfelton> | but there is no way to forcefully upgrade existing sessions in global.inc |
| [08:26:41] | <mrfelton> | where as secure pages does have a way to do that, although it doesn't seem to work very all either |
| [08:28:00] | <omega8cc> | mrfelton: as far as I remember, the simplest trick is to always initiate session over https and use redirect to http only for all unspecified paths where you don't need the session, but sometimes you need something smarter, I guess |
| [08:28:06] | * EclipseGc has joined #aegir |
| [08:29:14] | <mrfelton> | omega8cc: yeah, pretty much the conclusion we came to. But on this one site we just can't do that, because a part of the site that needs to be http for other reasons out of our control, and this part of the site has functionality on it that can cause a session to be initiated |
| [08:30:04] | <omega8cc> | yeah, so you need something better then |
| [08:30:29] | <omega8cc> | feel free to open an issue |
| [08:31:13] | <mrfelton> | omega8cc: ok, I'll do that. See if we can get some others to chime in with some possible solutions/workarounds |
| [08:31:31] | <omega8cc> | k |
| [08:31:43] | <mrfelton> | thanks |
| [08:37:14] | * lsolesen has quit (Quit: Leaving.) |
| [08:38:42] | <mrfelton> | omega8cc: k, another question relating my somewhat nonstandard use of Barracuda. We use Barracuda with remote servers which is not really supported, but it actually works very well. You just install Barracuda on the remote server, and then set it up as a remote server as normal in aegr, so that when you deploy to it from aegir, configs get copied over from the aegir server to the remote barracuda server which overwrites some of |
| [08:38:43] | <mrfelton> | configs that barracuda on the remote server installed. Basically, swapping ~/config/nginx.conf to point at server_remote instead of server_master. |
| [08:39:07] | <mrfelton> | We have been using it like this for a while now, and have upgraded barracuda several times with it set up this way without issue |
| [08:40:01] | <mrfelton> | This time however, aegir seems to be unable to sync some of the config files over to the remote barracuda server. Specifically, /var/aegir/config/includes/nginx_advanced_include.conf and /var/aegir/config/includes/nginx_simple_include.conf |
| [08:40:10] | <mrfelton> | and /var/aegir/config/includes/barracuda_log.txt |
| [08:40:35] | <mrfelton> | we get Operation not permitted from rsync which is unable to 'set times' on these files |
| [08:41:08] | <omega8cc> | hah |
| [08:41:10] | <mrfelton> | omega8cc just wondering if you knew if permissions had been changed on these config files in latest barracuda? |
| [08:41:24] | <omega8cc> | this is because we have converted them into symlinks |
| [08:41:29] | <mrfelton> | hmm, ok |
| [08:41:33] | <mrfelton> | I noticed that |
| [08:41:42] | <omega8cc> | just check ls -la /var/aegir/config/includes/ |
| [08:41:57] | <omega8cc> | so, we should revert something there |
| [08:42:02] | <omega8cc> | at least for master |
| [08:42:15] | <mrfelton> | omega8cc: yeah, I see that. So thats what is borking the rsync you think? |
| [08:42:23] | <omega8cc> | since we shouldn't break the stuff that way |
| [08:42:27] | <omega8cc> | yep |
| [08:43:35] | <omega8cc> | it is a bug, so please open an issue |
| [08:43:59] | <mrfelton> | ok, will do |
| [08:44:03] | <omega8cc> | thanks |
| [08:44:07] | <mrfelton> | on a side note, I get really confused by the difference between all these conf files… nginx_advanced_include.conf, nginx_octopus_include.conf, nginx_compact_include.conf, nginx_legacy_include.conf, nginx_modern_include.conf, nginx_simple_include.conf!! Are there any notes anywhere on the differences between all these files? |
| [08:44:23] | * gusaus has quit (Quit: gusaus) |
| [08:44:51] | <omega8cc> | we keep the old for backward compatibility, but symlinked to limit the confusion |
| [08:45:07] | <omega8cc> | there is an issue for that (where it is answered) |
| [08:45:23] | <omega8cc> | but, you know, we don't have real docs, yet |
| [08:45:30] | <mrfelton> | omega8cc: ok, so octopus and legacy are old ones? |
| [08:45:37] | <omega8cc> | no |
| [08:45:43] | <mrfelton> | ok, other way round is it? |
| [08:46:19] | <mrfelton> | just wondering which ones I can delete, so that I can get around this issue for now |
| [08:46:32] | <mrfelton> | i.e., need to get rid of the symlinks |
| [08:47:06] | <omega8cc> | see: http://drupal.org/node/1447780#comment-5637878 |
| [08:47:07] | <hefring> | http://drupal.org/node/1447780 => Best way to change "primary domain" from site.com to www.site.com => Octopus, Miscellaneous, normal, closed (fixed), 6 comments, 2 IRC mentions |
| [08:47:50] | <omega8cc> | you can safely delete all three symlinks if you don't have some very old vhosts referencing them |
| [08:47:54] | <mrfelton> | omega8cc: perfect, thanks! |
| [08:48:10] | <omega8cc> | np |
| [08:48:55] | <mrfelton> | omega8cc: I do have an old ghost referencing one of those. My guess is that if I reverify it, it should update, right?> |
| [08:50:09] | * gusaus has joined #aegir |
| [08:50:18] | <mrfelton> | yeah, it did :) |
| [08:50:30] | <omega8cc> | yep |
| [08:52:45] | <omega8cc> | mrfelton: we need to simplify this and anarcat did some good initial work on this: http://drupal.org/node/1622846 |
| [08:52:46] | <hefring> | http://drupal.org/node/1622846 => Nginx config include files should be merged into one => Provision, HTTP Service, normal, needs review, 0 comments, 1 IRC mention |
| [08:53:49] | <omega8cc> | we will still need separate copy for different ports used by php-fpm 5.2 and 5.3 in boa, but we will drop d5 support soon |
| [08:54:18] | <omega8cc> | I mean, by default, with option to still have it |
| [08:55:52] | <mrfelton> | omega8cc: thats great. Some cleanup is much needed in those config files I think. less duplication=good! |
| [08:56:55] | <omega8cc> | yeah, and we need drush 5 to properly support drupal 8, since our forked 4.6-dev no longer works, etc |
| [08:57:31] | <mrfelton> | omega8cc: Another thing I noticed. Now we include the Hosting queue runner, which is great. I have that running on a lot of my custom aegir servers. But, it does't seem to run properly here |
| [08:58:05] | <mrfelton> | sudo /etc/init.d/hosting-queue-runner status always reports that it's not running |
| [08:58:19] | <mrfelton> | and start does;t seem to actually start it |
| [08:59:17] | <omega8cc> | I didn't try it yet, as it is not really a good idea when you have a few octopus instances, so feature request, please :) |
| [09:00:07] | <mrfelton> | kk |
| [09:22:47] | * patcon has quit (Quit: patcon) |
| [09:37:10] | <mrfelton> | omega8cc: looks like _NGINX_EXTRA_CONF is no longer supported? |
| [09:37:33] | <mrfelton> | This was set in my ~/.barracuda.cnf file, but my options have not been compiled into nginx |
| [09:37:36] | * gboudrias has joined #aegir |
| [09:38:48] | <mrfelton> | strange… one server has nginx/0.8.54 whist another has nginx/1.3.0 |
| [09:38:58] | <mrfelton> | both upgraded to 2.0.3 just now |
| [09:41:11] | <mrfelton> | hmm, ok. Looks like one of these upgraded also stopped mid way. So I guess nginx install didn't happen. I'll try running again |
| [09:41:48] | <omega8cc> | it is because we don't install nginx from sources by default on squeeze and precise |
| [09:42:18] | <omega8cc> | you need to use _NGINX_FROM_SOURCES=YES |
| [09:42:41] | <omega8cc> | we use packages by default on squeeze and precise |
| [09:43:24] | <omega8cc> | I would suggest to review all available config options in the Barracuda script and adjust your cnf file |
| [09:43:41] | <omega8cc> | hmm |
| [09:44:10] | <mrfelton> | This is on Ubuntu 11.04 |
| [09:44:13] | <omega8cc> | we should probably automatically switch to _NGINX_FROM_SOURCES=YES when _NGINX_EXTRA_CONF is not empty |
| [09:44:59] | <omega8cc> | hmm, so it should work |
| [09:46:11] | <mrfelton> | omega8cc: yeah, it has now. I think I aborted the install process. It must have been when the installer was between installing 0.8.54 and upgrading it to 1.3.0. |
| [09:47:01] | <mrfelton> | yeah, all good now. correct nginx version and it has my addons |
